Re: [tip:x86/mm] [x86/mm/tlb] 209954cbc7: will-it-scale.per_thread_ops 13.2% regression
From: Oliver Sang
Date: Thu Nov 28 2024 - 20:44:27 EST
hi, Peter Zijlstra,
On Thu, Nov 28, 2024 at 05:21:28PM +0100, Peter Zijlstra wrote:
> On Thu, Nov 28, 2024 at 10:57:35PM +0800, kernel test robot wrote:
> >
> >
> > Hello,
> >
> > kernel test robot noticed a 13.2% regression of will-it-scale.per_thread_ops on:
> >
> >
> > commit: 209954cbc7d0ce1a190fc725d20ce303d74d2680 ("x86/mm/tlb: Update mm_cpumask lazily")
> > https://git.kernel.org/cgit/linux/kernel/git/tip/tip.git x86/mm
> >
[...]
> >
> >
> > The kernel config and materials to reproduce are available at:
> > https://download.01.org/0day-ci/archive/20241128/202411282207.6bd28eae-lkp@xxxxxxxxx
> >
> > =========================================================================================
> > compiler/cpufreq_governor/kconfig/mode/nr_task/rootfs/tbox_group/test/testcase:
> > gcc-12/performance/x86_64-rhel-9.4/thread/100%/debian-12-x86_64-20240206.cgz/lkp-skl-fpga01/tlb_flush2/will-it-scale
> >
> > commit:
> > 7e33001b8b ("x86/mm/tlb: Put cpumask_test_cpu() check in switch_mm_irqs_off() under CONFIG_DEBUG_VM")
> > 209954cbc7 ("x86/mm/tlb: Update mm_cpumask lazily")
>
> I got a bit lost in the actual jobs descriptions. Are you running this
> test with or without affinity? AFAICT will-it-scale itself defaults to
> being affine (-n is No Affinity).
with affinity. we don't change this default setting.
the command is
python3 ./runtest.py tlb_flush2 295 thread 0 0 104
the first '0' keeps affinity as 'yes'
print('Usage: runtest.py <testcase> <duration> <mode> <no_affinity> <smt> <threads...>', file=sys.stderr)
sys.exit(1)
...
affinity = " "
if int(sys.argv[4]) == 1:
affinity = " -n "